def game(): counter = 0 depth = 1 q = Quoridor() while True: q.print_board() print(q.valid_actions()) action = input('please input action code:') action = interpret(action) q.take_action(action) end, winner = q.check_end() if end: print('Game over, winner is ', winner) break q.alter() counter += 1 if counter > 12: depth = 2 action = minimax(q, depth) q.take_action(action) end, winner = q.check_end() if end: print('Game over, winner is ', winner) break q.alter()
def test_mcts_player(): from quoridor import Quoridor from policy_value_net import PolicyValueNet policy_value_net = PolicyValueNet(model_file=None) c_puct = 5 n_playout = 400 mcts_player = MCTSPlayer(policy_value_net.policy_value_fn, c_puct=c_puct, n_playout=n_playout, is_selfplay=1) q = Quoridor() action = mcts_player.get_action(q) q.take_action(action) q.alter() # print(q.valid_actions()) # try: action = mcts_player.get_action(q) # except: # print(q.valid_actions()) print('take action : ', action)